The problem with GPS is the illusion that it is generally regarded as being infallable by the user. Just as there are the ineviatable errors associated with maps (the map is only as good as the original accuracy of the survey together with the fact the map features can move e.g. rivers may change their course or dry up, cliff edges may receed over time, introduction of new roads and the plantation of new wooded areas with older wooded areas being cut or burnt etc) and compass errors (magnetic declination and local geomagnetism), the GPS also has associated errors with it. Some can be quite considerable if the GPS is not setup correctly.

Additional problems occour when aligning the computed GPS position in the WGS84 Lat/Lon coordinate system to the map wether it is a paper or electronic one. The electronic map may have even more errors associated with it because they may have been generated from the paper version. The paper to electronic conversion will again introduce some form of error i.e. the scanning and conversion from a bitmap image to a vector image. This may involve some considerable computer graphical editoring, which may again involve some form of human error.

General GPS errors can be helped along by ensuring the GPS almanac and Ephemeris details are upto date by leaving the GPS to download this information especially if the GPS has been turned of for a few weeks or months.

WAAS and EGNOS being turned on can help considerably with the GPS accuracy providing differential GPS to help eliminate mostly atmospheric and multipath broadcast signal errors. This will tend to improve GPS accuracy from around 10-20 metres (CEP) to 1-5 metres (CEP).

Ensure that the correct mapping datum and position format match the paper map which is in use. The conversion from GPS WGS84 to the local map such as UTM MGRS or OSGB grid involves some complex calculation as can seen from the following link
http://www.ordnancesurvey.co.uk/oswebsite/gps/docs/convertingcoordinatesEN.pdf

More modern higher sensitivity GPS models help with aquiring low power lock on to the satellite signals in wooded areas etc. Not all GPS's are created equal.

Ensure that the lasted firmwire for the GPS is also up to date. Some errors that might have introduced errors especially with coordinate transformations may well have been reported and consequently fixed.

If after ensuring that the GPS (with the latest firmware) has a good signal with a fresh almanac and Ephemeris and that WAAS or EGNOS is turned on and that the correct position format and datum are selected for the map you are using and you still find that the GPS is apparently giving the incorrect location then chances are the map itself will be wrong.

Generally I have not had a problem with my very basic Garmin eTrex when used in conjuction with the UK Ordnance Survey maps in providing a very accurate position. GPS accuracy in relation to the Ordnance Survey maps is a claimed 2 metres by Ordnance Survey.

I wouldn't trust the Garmin to point me in the right direction though. But that's what a compass is for.

What I do find interesting though is the idea that a GPS is somehow easier to use and therefore can be introduced to the navigator starting out before introducing map and compass work. Surely it must be the other way around.